CUED Publications database

Automated Computer Vision-Based Detection of Exposed Transverse Reinforcement for Post-Earthquake Safety Assessments

Brilakis, I and German, S and DesRoches, R Automated Computer Vision-Based Detection of Exposed Transverse Reinforcement for Post-Earthquake Safety Assessments. In: 6th World Conference on Structural Control and Monitoring, 2014-7-15 to 2014-7-17, Barcelona, Spain pp. 1436-1443.. (Unpublished)

Full text not available from this repository.


Current procedures in post-earthquake safety and structural assessment are performed by a triage team of structural engineers or certified inspectors. These procedures are inherently time-consuming and qualitative. Spalling has been accepted as an important indicator of significant damage to structural elements during an earthquake, and thus provides a sound springboard for a model in computer vision-based automated assessment procedures as is proposed in this research. A novel method that automatically detects regions of spalling on reinforced concrete columns and measures their properties in image data is desired. With respect to these efforts, the properties of spalled regions are determined based largely on the extent of exposure of reinforcement. Thus, this work focuses primarily on the efforts to detect the exposure of transverse reinforcement on reinforced concrete frame members. According to this method, the region of spalling is first isolated by way of a local entropy-based thresholding algorithm. Following this, the regions of exposed reinforcement (transverse and longitudinal) are identified by thresholding in the CMYK channels. Then, the Hough Transform is used to detect horizontal edges within the regions contained in each of the respective CMYK channel thresholded images. Then, all of the edges are combined, near-collinear lines are merged, and the result is output as a region of transverse reinforcement exposed on the element surface. The method was tested on a database of damaged RC column images collected after the 2010 Haiti earthquake, and comparison of the results with manual detection indicate the validity of the method.

Item Type: Conference or Workshop Item (UNSPECIFIED)
Divisions: Div D > Construction Engineering
Depositing User: Cron Job
Date Deposited: 17 Jul 2017 20:15
Last Modified: 19 Jul 2018 07:05